THE AMERICAN NA TURALIST. [Vol. XXXVII. 



assumption that it formed a part, or the whole, of a paired organ 

 attached to the sides of either the trunk or to the anterior ventral 

 part of the head. 



The probable mode of attachment of the joint, and its direc- 

 tion of movement, is indicated by the sharp slope of its anterior 

 end, the length of the process on its anterior lateral margin, 

 and by the shape and condition of the condyle. These condi- 

 tions indicate that the plate could not have been attached 

 directly to the sides of the body, for in that case it would have 

 to be directed diagonally backwards and the prominent anterior 

 lateral process would prevent that range of forward and back- 

 ward movement which the shape of the condyle clearly indicates 



The same argument would preclude the possibility of its being 

 attached directly to any of the marginal incisions in the oral 

 region of the ventral surface. In a word, there is no place 

 where there is any likelihood that such a structure could be 

 attached directly to the body, and there is no structure known 

 in the Ostracoderms with which it can be compared, if not with 

 the distal joint of the pectoral appendages of Bothriolepis or 

 Pterichthys. With this it corresponds fairly well except that it is 



W e therefore assume that the plate in question formed the 

 distal joint of an appendage something like that of Bothriolepis. 

 It now remains for us to find the plates that make up the armor 

 of the proximal portion of the appendages. 



Basal joint: I regard the structure shown in Figs. 4 and 5 

 as one of these plates. Two specimens of this plate are in the 

 collections at Dartmouth College; a nearly perfect one from 

 which this photograph was made, and a fragment of another. 

 One or more of these plates are also in the collection of the 

 Imp Acad, of St. Petersburg. The plate is 6 mm. long and 



